Week 5 Computer Polls

[BrutusBobcat]

The first one out today is Sagarin.

Sagarin ELO-CHESS (BCS):

1. Florida State
2. LSU
3. USC
4. Miami
5. Texas
6. Michigan State
7. Virginia Tech
8. Florida
9. Georgia
10. Wisconsin
11. Minnesota
12. Boston College
13. Georgia Tech
14. West Virginia
15. Arizona State
16. UCLA
17. Penn State
...
26. Ohio State

I'll post Billingsley, Colley and Massey when they come out (or someone else can if they beat me to it ).

Anderson-Hester and Wolfe do not yet have rankings for 2005.

The upside here is that the computer so far is giving some love to a few future opponents.

 

[Jeffcat]

florida state is #1......................ummmmmm no

 

[BuckeyeNation27]

I wonder why Sagarin is considered the joke of the college football world. every year he ranks Pac-10 the #1 conference, and now this.

 

[Steve19]

Sagarin's model still has the opening Bayesian estimates in it. In other words, his beginning estimates are still in, so this isn't his real ratings yet. I think it will be next week that enough games have been played to drop that. Until then, don't consider these ratings anything more than a guide.

 

[JXC]

How in the hell is Penn State ahead of ANYBODY? They've played NOBODY! NOBODY! What a complete joke. Miami, FL #4? I don't understand what we did to get ranked so low? Lose to a team, that even they still have in the top 5.

 

[BrutusBobcat]

Steve is correct. These rankings still use the pre-season seedings, but there are some things to look out for.

First, the teams we've beaten haven't done jack. SDSU got their first win against San Jose State last night. Miami has beaten Kent State (who lost to OU yesterday). Iowa beat Ball State and Northern Iowa.

Second, the teams still on our schedule look good to Sagarin's computer. Minny, MSU and PSU are teams that are going to wind up in our "W" column and it's good for us that they are ranked high.

All of these computer algorithms create their rankings based on a few simple criteria. Home losses are punished more than road losses, for example. Home wins over teams with losing records count less than road wins against teams with winning records, etc.

So Ohio State is #26 due to underachieving opponents and a home loss. Since margin of victory is NOT counted in any BCS computer poll, due to BCS policy, we get no more credit for a close loss at home to Texas than would LA-Lafayette if they'd played in Lafayette. Over the course of the season, this will all work out, but early on you get some funny results.

One way to look at computer polls is -- what have the teams achieved thus far? The fact that PSU is ranked so highly with a pretty lackluster schedule is a good indicator as to how bad everyone else's schedule is.

 

[sandgk]

Sagarin's model still has the opening Bayesian estimates in it. In other words, his beginning estimates are still in, so this isn't his real ratings yet. I think it will be next week that enough games have been played to drop that. Until then, don't consider these ratings anything more than a guide.

Actually, I think you can go farther. Till he drops the opening estimate and only uses current season data I simply would not even bother looking at the Sagarin.

 

[BrutusBobcat]

Massey is now out as well:

<pre>
<font color="#0080FF"> Chng Team W L Rating Sched Massey</font>

1 <a href="team.php?tm=193738">LSU</a> 1 0 3.163 2.01 ( 2) 6 IA
2 + 1 <a href="team.php?tm=193633">Florida St</a> 3 0 2.944 1.75 ( 6) 2 IA
3 + 6 <a href="team.php?tm=193659">Penn St</a> 4 0 2.721 1.61 ( 12) 28 IA
4 - 2 <a href="team.php?tm=193728">Florida</a> 4 0 2.583 1.51 ( 20) 7 IA
5 + 3 <a href="team.php?tm=193636">Miami FL</a> 2 1 2.558 1.95 ( 3) 5 IA
6 +16 <a href="team.php?tm=193674">Texas</a> 3 0 2.385 1.34 ( 37) 4 IA
7 + 9 <a href="team.php?tm=193717">Arizona St</a> 3 1 2.368 1.65 ( 11) 20 IA
8 - 3 <a href="team.php?tm=193666">Kansas</a> 3 0 2.342 1.28 ( 44) 53 IA
9 +15 <a href="team.php?tm=193643">Boston College</a> 3 1 2.293 1.60 ( 14) 17 IA
10 +44 <a href="team.php?tm=193729">Georgia</a> 4 0 2.235 1.21 ( 54) 8 IA

<font color="#0080FF"> Chng Team W L Rating Sched Massey</font>

11 +15 <a href="team.php?tm=193723">USC</a> 3 0 2.217 1.19 ( 58) 1 IA
12 +25 <a href="team.php?tm=193680">Cincinnati</a> 2 1 2.216 1.66 ( 9) 80 IA
13 +15 <a href="team.php?tm=193760">Texas-El Paso</a> 3 0 2.212 1.17 ( 61) 60 IA
14 + 7 <a href="team.php?tm=193656">Minnesota</a> 4 0 2.183 1.16 ( 66) 15 IA
15 +17 <a href="team.php?tm=193722">UCLA</a> 3 0 2.180 1.17 ( 64) 13 IA
16 +76 <a href="team.php?tm=193685">South Florida</a> 3 1 2.162 1.45 ( 25) 51 IA
17 +14 <a href="team.php?tm=193655">Michigan St</a> 4 0 2.151 1.10 ( 78) 11 IA
18 +40 <a href="team.php?tm=193735">Alabama</a> 4 0 2.140 1.08 ( 80) 21 IA
19 -13 <a href="team.php?tm=193665">Iowa St</a> 3 0 2.117 1.12 ( 75) 49 IA
20 +42 <a href="team.php?tm=193640">Virginia Tech</a> 4 0 2.107 1.06 ( 84) 3 IA
</pre>

Ohio State is at #32.

Note how high Penn State is ranked, and how low USC is ranked. Va Tech is also quite low. Again, the results are kind of nonsense right now, but some good Big Ten wins will bring OSU up.

 

[JXC]

Penn State, Cincinnti, Kansas, and South Florida hahahahahahaha.

 

[Sdgobucks]

What a fucking joke, and to think this is how we decide a national champion. Yeah a fucking playoff is ridiculous compared to this.

 

[BrutusMaximus]

Hey where the hell is Vandy? They are 4-0. These rankings are bullshit

 

[HabaneroBuck]

Since margin of victory is NOT counted in any BCS computer poll, due to BCS policy, we get no more credit for a close loss at home to Texas than would LA-Lafayette if they'd played in Lafayette. Over the course of the season, this will all work out, but early on you get some funny results.

That's too bad...USC cannot control who they play to a certain extent, but they have more than compensated by absolutely annihilating the spread in their games. Computers should factor in margin of victory in relation to some kind of spread factor to analyze whether a team is top-notch or not. Any team that goes 8-4 or so against the spread in a year is really playing at a high level. Of course, BCS computers are not going to factor something related to gambling into its equation. I just mention it because USC has covered three really high spreads in a row....and one computer has them at #11?

FWIW, Florida State has done quite well against the spread also this year, so their #1 is somewhat warranted from that standpoint.

 

[BrutusBobcat]

Hey where the hell is Vandy? They are 4-0. These rankings are bullshit

Colley Matrix had them at #5 last week.

 

[BrutusBobcat]

Colley Matrix is now out, and you are all going to LOVE this. Check out who in the flaming f*ck is #1....I'm speechless:

1. PENN ST
2. FLORIDA
3. FLORIDA ST
4. GEORGIA
5. MICH ST
6. MINNESOTA
7. VA TECH
8. USC
9. WISCONSIN
10. TEXAS
11. WVU
12. ALABAMA
13. UTEP
14. UCLA
15. VANDY
16. CALIFORNIA
17. VIRGINIA
18. GA TECH
19. WYOMING
20. MIAMI FL
21. TENNESSEE
22. INDIANA
23. IOWA ST
24. KANSAS ST
25. BOSTON COLL
......
34. Ohio State

Notable items here are: four B10 teams in the top 10, and 3 upcoming opponents in the top 6. By the end of October, this computer is going to be singing "Hang On Sloopy" and drawing Script Ohio on the screen...

Indiana at #22 is also good for some entertainment.

 

[BrutusBobcat]

Looks like the good Mr. Billingsley is working late this evening as well. Here's his top 25...maybe we'll like this one a little better.

1. USC
2. Georgia
3. Wisconsin
4. Virginia Tech
5. Texas
6. Florida State
7. Georgia Tech
8. Miami
9. Ohio State
10. Tennessee
11. Arizona State
12. Florida
13. LSU
14. California
15. Boise State
16. TCU
17. Michigan
18. Auburn
19. BC
20. WVU
21. Iowa State
22. Texas Tech
23. Utah
24. Louisville
25. Michigan State

I suspect that this computer used some kind of preseason seeding. It's hard to account for Michigan at 17 with two losses being ahead of Michigan State. The same argument goes for Boise State. My bet is that it uses last seasons final rankings as a pre-seed for this season to kick start the process.

Edit: As an editorial aside here, I think it's ridiculous that these computers can't factor in ANY kind of margin of victory here. That's the biggest bunch of nonsense in the BCS, IMO. If you get pummeled by MSU at home, that counts against you just as much as if they squeaked by you in OT. There's ways to do it without encouraging coaches to run up the score.